Members-Only
Recent Talks & Demos are for members only
You must be an AI Tinkerers active member to view these talks and demos.
Constrained Decoding: LLM Pixel Art
Exploring constrained decoding by masking token probabilities to force LLMs to generate pixel art grids, revealing token-level generation mechanics and structural control limitations.
Recently, I’ve been experimenting with low-level token generation using models like gpt-oss-20b and gemma-3-27b-it. I became particularly interested in constrained decoding, where you apply a mask to the model’s probability distribution at each generation step.
That led me to a fun question: what if you constrain the probabilities to a predefined set of “color tokens” and force the model to output them in a grid, essentially generating pixel art?
Predictably, the results were… chaotic. But the failures were fascinating, and I’m convinced the limitations were mostly due to my rushed implementation and time constraints.